Against Intellectual Monopoly

the road engine was being widely adopted and improved.
Once Selden's patent, No. 549,160, was awarded, it commanded royalties of
1.25 percent on the sale value of every automobile sold in the United States.
Selden's monopoly power had a dramatic impact on the future of the U.S.
automobile industry; it lead, de facto, to its reorganization under a much more oligopolistic structure than at the time Selden acquired its patent. We
learn the following from Stuart Graham's doctoral dissertation chapter:

    Selden had sold his patent 549,160 in 1899 to a syndicate for $10,000 and 20% of
any royalties. Early manufacturers who had originally seen the Selden patent as a
threat formed a cartel around the patent, the Association of Licensed Automobile
Manufacturers, which limited membership and licenses to manufacture under the
Selden paten t.12
    So, if you were wondering why the U.S. automobile industry developed so
quickly into the oligopoly we know and hate, a fair share of the roots lie
in bad "intellectual property" legislation and the intellectual monopoly it
created.
    In more recent days, Jerome Lemelson, who patented the "idea" of
machine vision and related data-identification techniques, has probably matched Selden in this dubious ranking. Bringing lawsuits eighteen
to thirty-nine years after initially filing for patents, it is estimated that
Lemelson's submarines collected on the order of $1.5 billion, primarily by
suing large end users such as Motorola and Ford. Although this is not an
example of IP-inefficiency, the Lemelson case and hundreds of less-known
ones are worth reflecting upon as a strange and socially inefficient consequence of our patent laws.
    Jerome Lemelson was most certainly a man of genius and quite dedicated
to his lifelong task of being an inventor. Anyone surfing the Web and reading
about his career on the hundreds of sites celebrating his genius will realize
that Lemelson invented and successfully patented dozens of interesting
devices and ideas. The problem is that he invented them only "so to speak,"
so that where the patent applications may have run hundreds of pages, the
useful information was quite generic and there is little evidence that the
ideas contained in them led to useful devices. Most of the ideas or devices
he invented never made it to the market, and those that did were developed
by someone else - as far as we can tell, without having benefited from the
original patented idea. Lemelson contented himself with either selling his
patents to producers interested in that line of business or suing them, when
someone else who was somewhere else, most often unaware of Lemelson's
discovery and patent, was producing a useful tool that could more or less be
related to the preexisting Lemelson patent. The issue here is not the oftendebated issue of whether Lemelson was or was not in good faith making
the claims he made. The issue is that what matters for social welfare are
copies of ideas, that is, ideas that materialize into goods and services that
are produced and that people use. Hence, Lemelson's contribution to social welfare was small, or even negative, as he cost at least $1.5 billions to firms
that were inventing by themselves and then producing goods and services
useful to consumers.33

    Submarine patents are an especially egregious problem because by the
time the claim is made, the cost of development is sunk, so there is no reason
for the submarine to allow the innovator even to cover his or her own costs.
The most recent extension of the patent term from seventeen to twenty
years measures the patent term from date of application rather than date
of award, which makes submarine patents more difficult. But as the case of
Rambus shows, submarine patents are still a significant social problem.
    Rambus is a "fabless" manufacturer of memory chips, meaning that it
does not actually manufacture chips but designs them, and sublets the actual
manufacture to other companies that have the large expensive "fabs" needed
to produce chips. More recently, as its own designs have not turned out to
be terribly successful, Rambus has switched to a new business model: trying
to collect license fees from other chip makers that have successful designs.
In the early 1990s, Rambus patented a number of memory chip-related
ideas. The most significant among these was the "idea" of including onchip phase-lock-loop (PLL) circuitry to control timing. It should